Where in-vivo imaging meets cytoarchitectonics: the relationship between cortical thickness and neuronal density measured with high-resolution [18F]flumazenil-PET.

MRI-based measurements of surface cortical thickness (SCT) have become a sensitive tool to quantify changes in cortical morphology. When comparing SCT to histological cortical thickness maps, a good correspondence can be found for many but not all human brain areas. Discrepancies especially arise in the sensory motor cortex, where histological cortical thickness is high, but SCT is very low. The aim of this study was to determine whether the relationship between cortical thickness and neuronal density is the same for different cytoarchitectonic areas throughout homo- and heterotypical isocortex. We assessed this relationship using high-resolution [(18)F]-labelled flumazenil (FMZ) PET and SCT-mapping. FMZ binds to the benzodiazepine GABA(A) receptor complex which is localized on axo-dendritic synapses, with a cortical distribution closely following the local density of neurons. SCT and voxelwise FMZ binding potential (BP(ND)) were assessed in ten healthy subjects. After partial volume correction, two subsets with a differential relationship between SCT and BP(ND) were identified: a fronto-parietal homotypical subset where neuronal density is relatively constant and mainly independent of SCT, and a subset comprising heterotypical and mainly temporal and occipital homotypical regions where neuronal density is negatively correlated with SCT. This is the first in-vivo study demonstrating a differential relationship between SCT, neuronal density and cytoarchitectonics in humans. These findings are of direct relevance for the correct interpretation of SCT-based morphometry studies, in that there is no simple relationship between apparent cortical thickness and neuronal density, here attributed to FMZ binding, holding for all cortical regions.